10 Enterprise Tech Trends for 2017 and Beyond
One word sums up this year in enterprise tech: clarity.
We learned that the emerging ecosystem of containers, microservices, cloud scalability, devops, application monitoring, and streaming analyticsis not a fad. It’s the future, already powering Silicon Valley’s and Seattle’s most advanced tech companies. Throw in machine learning and IoT and you have a comprehensive framework for the next phase of enterprise IT, with continuous improvement as its founding principle.
But what about next year? Well, when you know where today's enterprise tech stands, it's easier to look ahead. In that spirit, I offer my nine enterprise tech trends for the coming year and beyond (with no repeats from previous years!). Let’s start with the most obvious:
1. Advanced collaboration
After years of “business social networking” failures, Slack and its ballooning ecosystem have established chat-based collaboration as a first-order business application. Competitors abound, of course, from HipChat to Flock, and everyone wonders whether Microsoft Teams will be able to beat Slack at its own game -- particularly since Teams comes free with Office 365.
But if you ask me, it’s odd that simple chat-based collaboration has taken off, because the chat room metaphor has been around since IRC. Developers have engaged in a deeper form of collaboration from the time Linus Torvalds introduced Git as a way to organize revisions to the Linux kernel, with GitHub, Bitbucket, and GitLab offering today’s most popular Git implementations. Jon Udell and others have suggested that GitHub could provide the basis of all sorts of collaborations beyond code.
More exciting, though, is the notion that machine learning might enable collaborative platforms to gather people, resources, and data in an organization to form workgroups on the fly, which is an idea that Zorawar Biri Singh put forth in a recent InfoWorld interview. Silo-busting collaboration is the key to digital transformation, so machine intelligence to enable that seems like a prime opportunity in this space for years to come. Flock already shows flashes of it with its "magic search" feature.
2. Deep learning
AI and its subset machine learning owe much of their resurgence to the cloud’s ability to serve up gobs of compute, memory, and data, on which algorithms can gorge themselves and produce useful results quickly. That goes double for deep learning, a compute-intensive variety of machine learning that employs multiple layers of neural networks operating on the same problem at the same time for tasks ranging from image recognition to fraud detection to predictive analytics.
All the major clouds give customers the ability to crank up the horsepower required (including GPU processing) for deep learning, with Google’s TensorFlow in the lead, which is available both as a service on Google Cloud Platform and as an open source project. Over time, IBM’s Watson has gained deep learning abilities as well, now accessible to developers in the Bluemix cloud. New offerings from Microsoft Azure (Microsoft Cognizant Toolkit) and AWS (the MXNet framework plus the new Rekognition, Polly, and Lex services) help make this the hottest space around.
3. The incredible SQL comeback
For a few years it seemed like all we did was talk about NoSQL databases like MongoDB or Cassandra. The flexible data modeling and scale-out advantages of these sizzling new solutions were stunning. But guess what? SQL has learned to scale out, too -- that is, with products such as ClustrixDB, DeepSQL, MemSQL, and VoltDB, you can simply add commodity nodes rather than bulking up a database server. Plus, such cloud database-as-a-service offerings as Amazon Aurora and Google Cloud SQL make the scale-out problem moot.
At the same time, NoSQL databases are bending over backward to offer SQL interoperability. The fact is, if you have a lot of data then you want to be able to analyze it, and the popular analytics tools (not to mention their users) still demand SQL. NoSQL in its crazy varieties still offers tremendous potential, but SQL shows no sign of fading. Everyone predicts some grand unification of SQL and NoSQL. No one knows what practical form that will take.
4. The triumph of Kubernetes
We know what the future of applications looks like: microservices running in Docker containers on scalable cloud infrastructure. But when you break monolithic applications into microservices, you have a problem: You need to manage and orchestrate them. A few solutions have emerged to meet the challenge, including Apache Mesos, Docker Swarm, and Google Kubernetes.
It’s pretty clear at this point that Kubernetes has won, at least for now. Why shouldn’t it? After all, no company has had more experience running containers in production at scale than Google, using an internal system known as Borg, from which Kubernetes was derived. All the major clouds support Kubernetes, with CoreOS and Red Hat leading Kubernetes providers for both on-premises and cloud implementations. Add to those Heptio, a new startup formed by ex-Googler Craig McLuckie, co-founder of the Kubernetes project.
Kubernetes triumph may be short-lived, though, in part because we’re at such an early stage with containers. At the latest AWS re:Invent conference, for example, CTO Werner Vogels announced a slew of new container management and orchestration tools. Google will stick with Kubernetes for obvious reasons, but the cloud is where the action is and this contest is far from over. It’s just beginning.
5. Serverless computing
When you’re a developer, worrying about infrastructure, even the virtual kind, is a drag when you just want to concentrate on application logic and UI design. Serverless computing platforms take the industry’s long history of piling abstraction on top of abstraction to the next level so that such lowly concerns become a thing of the past. The serverless model also encourages developers to grab functions from a library and string them together, minimizing the amount of original code that needs to be written.
AWS Lambda is the best-known example of serverless computing, but other clouds have followed suit. Microsoft has Azure Functions and Google offers Cloud Functions. The startup Iron.io, which develops software for microservices workload management, also provides a serverless computing platform.
6. Custom cloud processors
Did you know that Amazon has a subsidiary that designs its own ARM processors for servers? Better known is Google’s foray into co-processing -- the Tensor Processing Unit specifically designed to accelerate machine learning. Plus, Microsoft has added FPGAs to its data centers to optimize particular applications such as machine learning and plans to offer tools to enable Azure customers to program FPGAs as well. At Amazon re:Invent last week, AWS introduced its own FPGAs in the form of a new F1instance type for EC2.
7. IoT interoperability
The established messaging protocol for IoT has long been MQTT, whose compact and efficient nature lends itself to low-power, relatively dumb devices. In May 2016, Google’s Nest subsidiary open sourced Thread, a mesh networking protocol that enables devices with more processing power to maintain peer-to-peer connections without relying on a hub.
The most interesting developments have emerged at the application layer. In October, the AllSeen Alliance merged with the Open Connectivity Foundation, which effectively unified the IoT software frameworks AllJoyn and IoTivity into a single open source project. More dramatically, at the Amazon re:Invent conference last week, AWS CEO Andy Jassy announced AWS Greengrass, a software core (and SDK) designed to run on IoT devices, enabling those devices to run AWS Lambda functions and connect securely to the AWS IoT platform. All the major public clouds now have IoT platforms, which are crucial to IoT progress, so you can expect Microsoft Azure, Google Cloud Platform, and IBM Cloud to deliver their own Greengrass-like offerings in 2017.
8. Hardware as a service
This one is kind of a sleeper. IDC predicts that in 2017, 10 percent of enterprises will begin exploring PC-as-a-service agreements with vendors. Reportedly, HP and Lenovo already have such rental programs in place. On the server side, Dell, HP, and Lenovo will begin offering Microsoft-managed servers preloaded with Azure Stack on a subscription basis. Oracle has the on-premises version of Oracle Cloud, dubbed Oracle Cloud Machine, that is offers via a “cloud-oriented subscription model.” Is this the end of capital investment in IT as we know it?
9. Big Data
Big data analysis is used in a number of ways to solve problems today. For example, police departments reduce crime without blanketing the city with patrol cars, by pinpointing likely crime hot spots at a given point in time based on real-time and historical data.
Build new data architectures to handle unstructured data and real-time input, which are disruptive changes today. The biggest inhibitor to enterprise IT adoption of big data analytics, however, isn't the data architecture; it's a lack of big data skills.
10. Python, Python, Python
OK, this one is a little silly. But each year, the ranks of Python programmers grow, with Python occupying the No. 4 position among all languages in the Tiobe Index. Python’s clean, English-like syntax has helped make it the most recommended first programming language.
People use it for everything, but in particular it has gained traction among data scientists. Moreover, Python has become the preferred language of devops engineers who write code to automate operations, and Python frameworks and IDEs continue to blossom. How devoted is the Python crowd? Here’s a clue: Python 3.6 will be released on Christmas Day